CRM/Utils/PDF/Document.php

   1 <?php
   2 /*
   3  +--------------------------------------------------------------------+
   4  | CiviCRM version 5                                                  |
   5  +--------------------------------------------------------------------+
   6  | Copyright CiviCRM LLC (c) 2004-2019                                |
   7  +--------------------------------------------------------------------+
   8  | This file is a part of CiviCRM.                                    |
   9  |                                                                    |
  10  | CiviCRM is free software; you can copy, modify, and distribute it  |
  11  | under the terms of the GNU Affero General Public License           |
  12  | Version 3, 19 November 2007 and the CiviCRM Licensing Exception.   |
  13  |                                                                    |
  14  | CiviCRM is distributed in the hope that it will be useful, but     |
  15  | WITHOUT ANY WARRANTY; without even the implied warranty of         |
  16  | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.               |
  17  | See the GNU Affero General Public License for more details.        |
  18  |                                                                    |
  19  | You should have received a copy of the GNU Affero General Public   |
  20  | License and the CiviCRM Licensing Exception along                  |
  21  | with this program; if not, contact CiviCRM LLC                     |
  22  | at info[AT]civicrm[DOT]org. If you have questions about the        |
  23  | GNU Affero General Public License or the licensing of CiviCRM,     |
  24  | see the CiviCRM license FAQ at http://civicrm.org/licensing        |
  25  +--------------------------------------------------------------------+
  26  */
  27
  28 /**
  29  *
  30  * @package CRM
  31  * @copyright CiviCRM LLC (c) 2004-2019
  32  */
  33
  34 require_once 'TbsZip/tbszip.php';
  35
  36 /**
  37  * Class CRM_Utils_PDF_Document.
  38  */
  39 class CRM_Utils_PDF_Document {
  40
  41   public static $ooxmlMap = [
  42     'docx' => [
  43       'dataFile' => 'word/document.xml',
  44       'startTag' => '<w:body>',
  45       'pageBreak' => '<w:p><w:pPr><w:pStyle w:val="Normal"/><w:rPr></w:rPr></w:pPr><w:r><w:rPr></w:rPr></w:r><w:r><w:br w:type="page"/></w:r></w:p>',
  46       'endTag' => '</w:body></w:document>',
  47     ],
  48     'odt' => [
  49       'dataFile' => 'content.xml',
  50       'startTag' => '<office:body>',
  51       'pageBreak' => '<text:p text:style-name="Standard"></text:p>',
  52       'endTag' => '</office:body></office:document-content>',
  53     ],
  54   ];
  55
  56   /**
  57    * Convert html to a Doc file.
  58    *
  59    * @param array $pages
  60    *   List of HTML snippets.
  61    * @param string $fileName
  62    *   The logical filename to return to client.
  63    *   Ex: "HelloWorld.odt".
  64    * @param array|int $format
  65    */
  66   public static function html2doc($pages, $fileName, $format = []) {
  67     if (is_array($format)) {
  68       // PDF Page Format parameters passed in - merge with defaults
  69       $format += CRM_Core_BAO_PdfFormat::getDefaultValues();
  70     }
  71     else {
  72       // PDF Page Format ID passed in
  73       $format = CRM_Core_BAO_PdfFormat::getById($format);
  74     }
  75     $paperSize = CRM_Core_BAO_PaperSize::getByName($format['paper_size']);
  76
  77     $metric = CRM_Core_BAO_PdfFormat::getValue('metric', $format);
  78     $pageStyle = [
  79       'orientation' => CRM_Core_BAO_PdfFormat::getValue('orientation', $format),
  80       'pageSizeW' => self::toTwip($paperSize['width'], $paperSize['metric']),
  81       'pageSizeH' => self::toTwip($paperSize['height'], $paperSize['metric']),
  82       'marginTop' => self::toTwip(CRM_Core_BAO_PdfFormat::getValue('margin_top', $format), $metric),
  83       'marginRight' => self::toTwip(CRM_Core_BAO_PdfFormat::getValue('margin_right', $format), $metric),
  84       'marginBottom' => self::toTwip(CRM_Core_BAO_PdfFormat::getValue('margin_bottom', $format), $metric),
  85       'marginLeft' => self::toTwip(CRM_Core_BAO_PdfFormat::getValue('margin_left', $format), $metric),
  86     ];
  87
  88     $ext = pathinfo($fileName, PATHINFO_EXTENSION);
  89
  90     $phpWord = new \PhpOffice\PhpWord\PhpWord();
  91
  92     $phpWord->getDocInfo()
  93       ->setCreator(CRM_Core_DAO::getFieldValue('CRM_Contact_BAO_Contact', CRM_Core_Session::getLoggedInContactID(), 'display_name'));
  94
  95     foreach ((array) $pages as $page => $html) {
  96       $section = $phpWord->addSection($pageStyle + ['breakType' => 'nextPage']);
  97       \PhpOffice\PhpWord\Shared\Html::addHtml($section, $html);
  98     }
  99
 100     self::printDoc($phpWord, $ext, $fileName);
 101   }
 102
 103   /**
 104    * @param object|string $phpWord
 105    * @param string $ext
 106    *   File extension/type.
 107    *   Ex: docx, odt, html.
 108    * @param string $fileName
 109    *   The logical filename to return to client.
 110    *   Ex: "HelloWorld.odt".
 111    *   Alternatively, a full path of a file to display. This seems sketchy.
 112    *   Ex: "/var/lib/data/HelloWorld.odt".
 113    */
 114   public static function printDoc($phpWord, $ext, $fileName) {
 115     $formats = [
 116       'docx' => 'Word2007',
 117       'odt' => 'ODText',
 118       'html' => 'HTML',
 119       // todo
 120       'pdf' => 'PDF',
 121     ];
 122
 123     if (realpath($fileName)) {
 124       $phpWord = \PhpOffice\PhpWord\IOFactory::load($fileName, $formats[$ext]);
 125     }
 126
 127     //CRM-20015
 128     \PhpOffice\PhpWord\Settings::setOutputEscapingEnabled(TRUE);
 129     $objWriter = \PhpOffice\PhpWord\IOFactory::createWriter($phpWord, $formats[$ext]);
 130
 131     CRM_Utils_System::setHttpHeader('Content-Type', "application/$ext");
 132     CRM_Utils_System::setHttpHeader('Content-Disposition', 'attachment; filename="' . $fileName . '"');
 133     $objWriter->save("php://output");
 134   }
 135
 136   /**
 137    * @param $value
 138    * @param $metric
 139    * @return int
 140    */
 141   public static function toTwip($value, $metric) {
 142     $point = CRM_Utils_PDF_Utils::convertMetric($value, $metric, 'pt');
 143     return \PhpOffice\PhpWord\Shared\Converter::pointToTwip($point);
 144   }
 145
 146   /**
 147    * @param array $path  docx/odt file path
 148    * @param string $type  File type
 149    *
 150    * @return array
 151    *   Return extracted content of document in HTML and document type
 152    */
 153   public static function docReader($path, $type) {
 154     $type = array_search($type, CRM_Core_SelectValues::documentApplicationType());
 155     $fileType = ($type == 'docx') ? 'Word2007' : 'ODText';
 156
 157     $phpWord = \PhpOffice\PhpWord\IOFactory::load($path, $fileType);
 158     $phpWordHTML = new \PhpOffice\PhpWord\Writer\HTML($phpWord);
 159
 160     // return the html content for tokenreplacment and eventually used for document download
 161     return [$phpWordHTML->getWriterPart('Body')->write(), $type];
 162   }
 163
 164   /**
 165    * Extract content of docx/odt file
 166    *
 167    * @param string $filePath  Document file path
 168    * @param string $docType  File type of document
 169    *
 170    * @return array
 171    *   [string, clsTbsZip]
 172    */
 173   public static function unzipDoc($filePath, $docType) {
 174     $dataFile = self::$ooxmlMap[$docType]['dataFile'];
 175
 176     $zip = new clsTbsZip();
 177     $zip->Open($filePath);
 178     $content = $zip->FileRead($dataFile);
 179
 180     return [$content, $zip];
 181   }
 182
 183   /**
 184    * Modify contents of docx/odt file(s) and later merged into one final document
 185    *
 186    * @param array $contents
 187    *   Content of formatted/token-replaced document.
 188    *   List of HTML snippets.
 189    * @param string $fileName
 190    *   The logical filename to return to client.
 191    *   Ex: "HelloWorld.odt".
 192    * @param string $docType
 193    *   Document type e.g. odt/docx
 194    * @param clsTbsZip $zip
 195    *   Zip archive
 196    * @param bool $returnFinalContent
 197    *   Return the content of file document as a string used in unit test
 198    *
 199    * @return string
 200    */
 201   public static function printDocuments($contents, $fileName, $docType, $zip, $returnFinalContent = FALSE) {
 202     $dataMap = self::$ooxmlMap[$docType];
 203
 204     $finalContent = $zip->FileRead($dataMap['dataFile']);
 205
 206     // token-replaced document contents of each contact will be merged into final document
 207     foreach ($contents as $key => $content) {
 208       if ($key == 0) {
 209         $finalContent = $content;
 210         continue;
 211       }
 212
 213       // 1. fetch the start position of document body
 214       // 2. later fetch only the body part starting from position $start
 215       // 3. replace closing body tag with pageBreak
 216       // 4. append the $content to the finalContent
 217       $start = strpos($content, $dataMap['startTag']);
 218       $content = substr($content, $start);
 219       $content = str_replace($dataMap['startTag'], $dataMap['pageBreak'], $content);
 220       $finalContent = str_replace($dataMap['endTag'], $content, $finalContent);
 221     }
 222
 223     if ($returnFinalContent) {
 224       return $finalContent;
 225     }
 226
 227     // Replace the loaded document file content located at $filePath with $finaContent
 228     $zip->FileReplace($dataMap['dataFile'], $finalContent, TBSZIP_STRING);
 229
 230     $zip->Flush(TBSZIP_DOWNLOAD, $fileName);
 231   }
 232
 233 }